# Hey Chat, Can You Teach Me? Structuring Socratic Dialogue for Human Learning in the Wild

arXiv: https://arxiv.org/abs/2606.11744 · HTML: https://mlsanity.com/paper/2606.11744/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 32 references that are in the map.
- First posted on arXiv 10 Jun 2026.

## Builds on (most cited first, 19 of 19)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World](https://mlsanity.com/paper/1703.06907/) (2017)
- [LLMs Get Lost In Multi-Turn Conversation](https://mlsanity.com/paper/2505.06120/) (2025)
- [Deep Knowledge Tracing](https://mlsanity.com/paper/1506.05908/) (2015)
- [Automatic Generation of Socratic Subquestions for Teaching Math Word Problems](https://mlsanity.com/paper/2211.12835/) (2022)
- [Training LLM-based Tutors to Improve Student Learning Outcomes in Dialogues](https://mlsanity.com/paper/2503.06424/) (2025)
- [Exploiting Cognitive Structure for Adaptive Learning](https://mlsanity.com/paper/1905.12470/) (2019)
- [LearnLM: Improving Gemini for Learning](https://mlsanity.com/paper/2412.16429/) (2024)
- [Enhancing Critical Thinking in Education by means of a Socratic Chatbot](https://mlsanity.com/paper/2409.05511/) (2024)
- [From Problem-Solving to Teaching Problem-Solving: Aligning LLMs with Pedagogy using Reinforcement Learning](https://mlsanity.com/paper/2505.15607/) (2025)
- [Evaluating Gemini in an arena for learning](https://mlsanity.com/paper/2505.24477/) (2025)
- [A Comprehensive Review of AI-based Intelligent Tutoring Systems: Applications and Challenges](https://mlsanity.com/paper/2507.18882/) (2025)
- [If the Sources Could Talk: Evaluating Large Language Models for Research Assistance in History](https://mlsanity.com/paper/2310.10808/) (2023)
- [Personalized Exercise Recommendation with Semantically-Grounded Knowledge Tracing](https://mlsanity.com/paper/2507.11060/) (2025)
- [FoundationalASSIST: An Educational Dataset for Foundational Knowledge Tracing and Pedagogical Grounding of LLMs](https://mlsanity.com/paper/2602.00070/) (2026)
- [A Training-Free Large Reasoning Model-based Knowledge Tracing Framework for Unified Prediction and Prescription](https://mlsanity.com/paper/2601.01708/) (2026)
- [EduQate: Generating Adaptive Curricula through RMABs in Education Settings](https://mlsanity.com/paper/2406.14122/) (2024)
- [UCO: A Multi-Turn Interactive Reinforcement Learning Method for Adaptive Teaching with Large Language Models](https://mlsanity.com/paper/2511.08873/) (2025)

## Built on by (most cited first, 1 of 1)

- [Do LLM Tutors Actually Teach? A Process-Centered Synthesis of the Tutoring Loop](https://mlsanity.com/paper/2610.15692/) (2026)

## Abstract (arXiv)

Large language models are now widely used for everyday learning, but the underlying interactions are typically unstructured chats rather than following a curriculum. Unlike formal online learning systems, these interactions carry no prior record of the student, so any estimate of what the student already knows must be inferred from the dialogue itself. We show that this gap is not closed by scaling models alone. Frontier and education-tuned LLMs perform poorly when asked to tutor a student over an extended session, because doing so requires three things at once. The tutor must sequence a curriculum, conduct Socratic dialogue, and infer the student's knowledge state from that dialogue. We propose separating these responsibilities. Given a student query, our system constructs a prerequisite knowledge graph in which subtopics are nodes and dependencies are edges, and frames tutoring as deciding which node to teach next and how many dialogue turns to spend on it before moving on. A lightweight PPO policy handles this sequencing decision, while an LLM conducts the Socratic exchange at the chosen node and returns a signal of student progress. Across held-out STEM and non-STEM topics, our PPO-paired tutor outperforms heuristic baselines, frontier general-purpose models, and a model specialised for Socratic dialogue: on both the rate at which students reach full curriculum mastery and the number of turns required. Explicit curriculum structure delivers gains that scaling the underlying model does not.